FORUM Swede Midge (Diptera: ), Ten Years of Invasion of Crucifer Crops in North America

MAO CHEN,1 ANTHONY M. SHELTON,1,2 REBECCA H. HALLETT,3 CHRISTINE A. HOEPTING,4 5 1 JULIE R. KIKKERT, AND PING WANG

J. Econ. Entomol. 104(3): 709Ð716 (2011); DOI: 10.1603/EC10397 ABSTRACT The Swede midge, nasturtii Kieffer (Diptera: Cecidomyiidae), a common pest in Europe, is a newly invasive pest in North America that constitutes a major threat to cruciferous vegetable and Þeld crops. Since its Þrst identiÞcation in , Canada, in 2000, it has rapidly spread to 65 counties in the provinces of Ontario and and has recently been found in canola (one of two cultivars of rapeseed, Brassica napus L. and Brassica campestris L.) in the central Prairie region where the majority of CanadaÕs 6.5 million ha (16 million acres) of canola is grown. The Þrst detection of Swede midge in the United States was in 2004 in (Brassica oleracea L.), but it has now been found in four additional states. Here, we review the biology of Swede midge, its host plant range, distribution, economic impact, pest status, and management strategies. We provide insight into this insectÕs future potential to become an endemic pest of brassica crops in North America. We also proposed research needed to develop tactics for handling this invasive pest in brassica crops.

KEY WORDS Contarinia nasturtii,invasivespecies,NorthAmerica,integratedpestmanagement

Cruciferous plants (BrassicaceaeϭCruciferae) form a cosmetic pest of headed cabbage (Shelton et al. 1998). diverse set of wild and cultivated plants encompassing However, by far the most effort has been devoted to Ͼ375 genera (Jones and Luchsinger 1979). In tem- developing thresholds for foliar-feeding Lepidoptera perate regions, cultivated crucifers include vegetables (e.g., Shelton et al. 1983), and these continue to be and oil-bearing Þeld crops in the genera Brassica L. reÞned for use in extension-oriented publications and Raphanus L. Regardless of the crop use, cultivated (OMAFRA 1996, Eastman et al. 2005, Reiners and crucifers can be attacked by 50Ð60 insect species, of Petzoldt 2010). In the New York cabbage IPM pro- which Ϸ20 are major pests (Finch and Thompson gram, early adoption of thresholds for control of Lep- 1992). The major insect species attacking crucifers idoptera led to a 49% decrease in insecticide use and belong to the orders Lepidoptera, Coleoptera, a54%increaseintheeffectivenessofeachinsecticide Hemiptera, Diptera, and Thysanoptera. All different application (Andaloro et al. 1983). Because of the parts and stages of crucifer plants can be damaged by success of these thresholds, it has been estimated that belonging to one or more of these orders. Ͼ80% of all cabbage growers in New York use major Considerable efforts have been made to develop components of the IPM guidelines for insect manage- components that can be used in integrated pest man- ment on cabbage (Reiners and Petzoldt 2010). We agement (IPM) programs for the most important cru- believe there are similar high adoption rates of IPM cifer pests on high-value cruciferous vegetables. For practices for cruciferous vegetables in other states and example, in cabbage (Brassica oleracea L.) degree-day in Canada. However, the advances made to such IPM models can be used to predict ßights of adult cabbage programs over the last several decades are at risk maggot, Delia radicum (L.), to help time early season because of a new invasive insect pest that has resulted insecticide treatments and cultural management prac- in some growers reverting to weekly spray programs. tices (e.g., Jyoti et al. 2003). However, host plant Swede midge, Contarinia nasturtii Kieffer (Diptera: resistance has been the most reliable practice for con- Cecidomyiidae), is a gall-forming pest of cruciferous trolling onion thrips, Thrips tabaci Lindeman, a serious plants that is common and endemic in Europe (Harris 1966) and southwestern Asia (Barnes 1946, Darvas et al. 2000). Host plants of Swede midge include most 1 Department of Entomology, New York State Agricultural Exper- cultivated cruciferous vegetables (e.g., , cab- imental Station, Cornell University, Geneva, NY 14456. 2 Corresponding author, e-mail: [email protected]. bage, caulißower, Brussels sprouts, and kale [all forms 3 School of Environmental Sciences, University of Guelph, Guelph, of B. oleracea]), cruciferous weeds, and canola (one of ON, Canada NIG 2W1. two cultivars of rapeseed, Brassica napus L. and Bras- 4 Cornell University Cooperative Extension Vegetable Program, sica campestris L.) (Barnes 1946, Stokes 1953a,b; Dar- 12690 Route 31, Albion, NY 14411. 5 Cornell University Cooperative Extension Vegetable Program, vas et al. 2000; Hallett 2007; Chen et al. 2009b). Plant 480 N. Main St., Canandaigua, NY 14424. damage results from larval feeding which causes mis-

0022-0493/11/0709Ð0716$04.00/0 ᭧ 2011 Entomological Society of America 710 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3 shapen plants with twisted stems, crumpled leaves, pest and to present a concise review of its biology and swollen growing tips, multiple heads, and the forma- management. tion of galls on leaves and ßowers (Bardner et al. 1971, Allen et al. 2008, http://www.nysaes.cornell.edu/ent/ Swede Midge Description and Life Cycle swedemidge/biology.html), all of which can severely reduce product quantity, quality and marketability. In In Ontario and Quebec, Canada, where Swede many parts of Europe, Swede midge is considered a midge were Þrst detected and remain the locations in major pest with frequent crop losses in spite of regular North America where they are most abundant, adults chemical treatments (den Ouden et al. 1987). Before (1.5Ð2 mm) emerge in mid- to late May, after over- 2000, it was not known to occur in North America, wintering in the soil in larval cocoons (Hallett et al. although damage symptoms typical of Swede midge 2007, 2009b; Corlay and Boivin 2008). Mating occurs on crucifer crops were observed in Ontario, Canada, soon after emergence and females then begin to look in 1996 (Hallett and Heal 2001). for suitable hosts. Eggs are laid in clusters of two to 50 In June 2000, the Þrst Swede midge in North Amer- eggs on meristematic tissue, young leaves, or ßowers ica was captured using yellow sticky traps in Ontario, of their host plants (Barnes 1946), with each female Canada (Hallett and Heal 2001). Subsequently, Ca- laying Ϸ100 eggs during her short lifetime (1Ð5 d). nadian researchers launched a Swede midge survey in Eggs are small (Ϸ0.3 mm) with color changing from cruciferous crops, and good evidence was found that transparent to a creamy white during the egg devel- Swede midge occurred in nine counties in Ontario and opmental period. The larval stage is the only life stage one county in Quebec in 2001. The number of infested that can damage cruciferous plants. Swede midge lar- counties in these provinces increased from 18 in 2005 vae have a well-developed mouth apparatus consisting to 49 in 2006 and 65 in 2008 (CFIA 2008). In 2007 and of the labrum, labium, mandibles, and maxillae. The 2008, additional Canadian records were found in Nova exact feeding mechanism of Swede midge larvae is not Scotia, Prince Edward Island, Manitoba, and Saskatch- clear; however, it is believed to be similar to other ewan (CFIA 2008). Detection was aided by the use of Cecidomyiidae species, i.e., larvae inject salivary ßuids pheromone traps that became available in 2004 (Hill- into the plant via highly specialized mandibles, and bur et al. 2005). In the United States, Swede midge was these substances react with plant cell components and cause physiological alterations in the plant, which al- Þrst reported in Niagara County, NY, in September low the to suck liquids through its mouth (Hatch- 2004, although symptomatic plants had been observed ett et al. 1990). Larvae are initially Ϸ0.3 mm in length in the Þeld for several years (Kikkert et al. 2006). By and reach a Þnal size of 3Ð4 mm. They are transparent the end of 2005, Swede midge was detected in Þve and become increasingly more yellow with age and are major cabbage-producing counties (Erie, Genesee, lemon-yellow at maturity. The larvae feed for 7Ð21 d Monroe, Orleans, and Wyoming) in New York state. before they drop to the ground to pupate. During An additional 13 counties in New York and one each periods of drought, larvae in the soil may remain dor- in (Hampshire), (Sussex), mant, with growth resuming after rainfall or irrigation and (New Haven) have been reported (Readshaw 1966). having Swede midge based on molecular and mor- Seasonal development of Swede midge was studied phological identiÞcations of adults caught in phero- in Ontario and Quebec, Canada, by using pheromone mone traps (Chen et al. 2007, 2009b). To date, Swede traps and emergence cages (Goodfellow 2005, Corlay midge has been detected in a total of 26 counties in and Boivin 2008). In Quebec, three to four generations New York, in addition to the three in surrounding of Swede midge were detected in 2004 and 2005 (Cor- states. An additional Swede midge infestation in Ot- lay and Boivin 2008). Based on MidgEmerge, a new tawa Co., OH, was conÞrmed in late 2009 based on predictive model developed using DYMEX and On- Swede midge mitochondrial cytochrome oxidase sub- tario and Quebec data on Swede midge adults and unit I gene (COI)sequencedata(M.C.etal.,unpub- larvae, Hallett et al. (2009b) determined that cur- lished data). rently there are four overlapping generations of In the past 10 yr, Swede midge infestations in North Swede midge per year and two emergence pheno- America have spread rapidly and created intense re- types in North America with the last adult ßights sponses from vegetable and canola farmers, especially occurring in late September to early October. in some areas of Canada. Crop losses due to Swede midge on some farms in Ontario, CA, have been re- ported to be as much as 85% (Hallett and Heal 2001). Swede Midge Detection and Identification Researchers in the United States and Canada have It is difÞcult to detect Swede midge in the Þeld been actively testing various strategies to control because adults are short lived; pupae burrow into the Swede midge and prevent its further spread in North soil; and the small larvae feed cryptically, near growing America. To understand its pest status and manage- tips of plants and between tightly compressed leaves ment options, here we summarize present knowledge and petioles. Depending upon sampling time, larvae on its biology and control strategies (e.g., chemical, may be absent from plants exhibiting damage symp- cultural control, and biological control) in North toms, making positive determination of an infestation America. This is done to alert the larger entomological difÞcult. Characteristic damage symptoms on plants community in North America of this newly invasive (e.g., swollen, distorted or twisted young shoots and June 2011 CHEN ET AL.: SWEDE MIDGE IN NORTH AMERICA 711 leaf stalks, blind heads due to death of the meristem, speciÞc primers. This PCR-based method has been brown scars on petioles or stems, and the presence of modiÞed for more rapid identiÞcation of the Swede secondary bacterial rots) are often used as visual cues midge and successfully used for identiÞcation of for the presence of Swede midge in Þelds (Allen et al., Swede midge specimens collected in the United States 2008, http://www.nysaes.cornell.edu/ent/swedemidge). (Kikkert et al. 2006; Chen et al. 2009b). However, these damage symptoms may easily be mis- Although the Swede midge pheromone blend has taken for common physiological (heat or frost stress) been reported to have high species speciÞcity in Eu- or nutritional problems (molybdenum deÞciency), rope (Hillbur et al. 2005), based on our experience, mechanical damage or other insect feeding. Other other midge species are regularly captured by the detection tools, such as sticky cards, incandescent and pheromone blend in New York state. Thus, capture in blacklight traps, Þne mesh emergence tents, and apheromonetrapandtheaccompanyingmolecular sweep netting have been proposed, but they are not identiÞcation is required for accurate identiÞcation of very practical due to low efÞciency and indiscrim- the insect. inating captures (Hallett et al. 2007). Hillbur et al. (2005) found that a 1:2:0.02 mixture of (2S,9S)- Host Range, Distribution, and Economic Impact diacetoxyundecane:(2S,10S)-diacetoxyundecane:(S)- 2-acetoxyundecane extracted from the femaleÕs ovi- The Swede midge reportedly feeds on all cultivars positor could effectively attract Swede midge males of B. oleracea, B. napus, Brascica rapa L., and Raphanus under laboratory and Þeld conditions with very high sativus L., and many common cruciferous weed spe- species speciÞcity. The Swede midge sex pheromone cies, such as mustard (Brassica spp.), wild blend maintained its attractiveness to Swede midge (Raphanus raphanistrum L.), shepherdÕs purse [Cap- males for only 2 wk in Þelds if cotton dispensers were sella bursa-pastoris (L.) Medik.], Þeld pepperweed used; however, if polyethylene dispensers were used, [Lepidium campestre (L.) R. Br.], Þeld pennycress its attractiveness could extend to Ͼ6wkinSwitzerland ( L.), and yellow rocket (Barbarea vul- (Boddum et al. 2009). However, under Ontario Þeld garis R. Br.) (Stokes 1953a,b; Hallett 2007; Chen et al. conditions, cotton wicks and polyethylene dispensers 2009b). A comprehensive list of Swede midge host were found to be attractive for only 2 and 4 wk, plant species is available on the Cornell University respectively (R.H.H. et al. unpublished data). Avail- Swede midge website (http://www.nysaes.cornell. ability of the pheromone blend greatly increases the edu/ent/swedemidge). Although Swede midge has a ease and efÞcacy of Swede midge detection in Þelds. wide range of host plants, different host species show Hallett et al. (2007) reported that pheromone traps different susceptibility to Swede midge damage (Hal- captured signiÞcantly more Swede midge than emer- lett 2007). In a choice test, Chen et al. (2009b) found gence traps, light traps and sticky cards, and subse- that Swede midge signiÞcantly preferred caulißower quent work determined white Jackson traps (Scentry plants for oviposition over six common cruciferous Biologicals, Inc., Billings, MT) to be more effective weed species. It is unclear what role cruciferous weeds than other trap types and colors (R.H.H. et al., un- play in terms of Swede midge occurrence and popu- published data). lation dynamics, but cruciferous weeds may serve as a Accurate identiÞcation of Swede midge has proven reservoir to maintain Swede midge populations in difÞcult, especially based on its morphological char- agroecosystems when no suitable crop hosts are avail- acteristics. Species identiÞcation based on morpho- able (Chen et al. 2009b). logical characters depends on examination of adult Temperature and moisture are considered to be the specimens. Swede midge wing venation is reduced two most important factors responsible for Swede with a straight radial vein and cubital fork in the midge population distribution and growth (Readshaw middle third of the wing, whereas cross veins are 1961). Using a bioclimatic model, Olfert et al. (2006) absent; adults have long, Þliform antennae with 12 determined that all Canadian provinces have suitable ßagellomeres on each antenna; female antennal seg- conditions for Swede midge establishment, with some ments are cylindrical with each ßagellomere Ϸ2times areas being very favorable for Swede midge popula- longer than wide, whereas each ßagellomere of a male tion growth. A revised version of that model (Mika et antenna is divided into two separate nodes surrounded al. 2008) indicated potential for establishment south to by a thread like looped sensillum (Harris 1966, Hallett Georgia, west to Colorado, Wyoming, Montana, and and Heal 2001). However, it is often very difÞcult to parts of Washington and Oregon. Both models used observe these characteristics in Þeld-collected speci- only natural rainfall. When irrigation was included in mens captured on sticky card traps or sticky surfaces the model, some western states, including California, in pheromone traps. In 2004, Frey et al. (2004) de- amajorbroccoli-producingstate,becameincludedin veloped a polymerase chain reaction (PCR)-based the potential range of infestation. With such an ex- molecular method for the Swede midge identiÞcation. tensive potential range, the introduction of Swede This PCR-based Swede midge identiÞcation method midge into additional areas could result in the wide- involves two rounds of PCR reactions, with the Þrst spread distribution of Swede midge throughout North PCR reaction to amplify a 440-bp fragment of the COI America. Under multiple climate change scenarios, gene from the specimens and with the second PCR Mika et al. (2008) project an increase to Þve gener- reaction (i.e., nested PCR) to amplify a 285-bp Swede ations per year, as well as substantial increases in areas midge-speciÞc fragment with a pair of Swede midge- designated as favorable or very favorable for Swede 712 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3 midge establishment and population growth in North etables. Initial Þeld efÞcacy trials in Ontario, Canada, America, with the Canadian prairies, northern Ontario in 2001Ð2002, focused on the pyrethroids, due to their and Quebec, and the eastern seaboard and midwest- use in Europe for Swede midge control, and led to ern United States becoming most vulnerable to Swede registration of ␭-cyhalothrin for Swede midge control midge establishment. in Canada in 2003 (Hallett et al. 2009a). Twenty in- Swede midge can be a devastating pest to crucifer- secticides belonging to 12 different classes were eval- ous crops because it damages growing tips of the uated using foliar spray, soil drenches, or seed treat- plants, which makes it impossible for plants to recover. ment in U.S. laboratory trials against Swede midge In addition, a wide range of host plants, multiple in- adults and larvae (Wu et al. 2006). Acephate, ␭-cyh- vasion paths (such as soil, transplant movement, wind, alothrin, acetamiprid, chlorpyrifos and methomyl adult ßight), cryptic feeding behavior, short-lived were effective as foliar spray insecticides against lar- adults, and a lack of effective control methods allow vae, and except for acetamiprid, they were also effec- Swede midge populations to thrive and cause signif- tive on adults. Acetamiprid, imidacloprid, and thia- icant losses. Crop losses on some Canadian farms due methoxam were very effective when applied as soil to Swede midge have been reported to be as much as drench and provided 100% control of Swede midge 85% (Hallett and Heal 2001). In some areas of Europe, larvae for 7 wk. Clothianidin and thiamethoxam pro- Swede midge infestations may account for 100% loss of vided 100% control when used as seed treatments. the crop, despite treatment with insecticides. Because the movement of vegetable seedlings could Crucifers are an extremely important group of veg- be an important reason for the rapid spread of Swede etables grown throughout the United States. The main midge in North America, Chen et al. (2007) evaluated fresh market vegetable crucifers (broccoli, cabbage, the efÞcacy of foliar sprays of acetamiprid on Swede and caulißower) were grown on nearly 121,406 ha midge on caulißower seedlings. Acetamiprid caused (300,000 acres) nationwide and had a total value of 99.5, 100, and 99.8% mortality of larvae when cauli- US$1.3 billion in 2009 (USDAÐERS 2010). In addition ßower seedlings were sprayed before inoculation with to these cruciferous vegetables, canola (rapeseed) is Swede midge 0 and4dafterinoculation,respectively. another host for Swede midge and is annually grown The efÞcacy of acetamiprid was reduced to 69.9% on 445,155 ha (Ͼ1.1 million acres) in the United when seedlings were sprayed at8daftertheinocu- States. More than 6,474,981 ha (16 million acres) of lation because Swede midge larvae could successfully canola is grown in Canada (Statistics Canada 2010). In pupate and emerge after the spray. Thus, acetamiprid 2005, Swede midge was Þrst reported to have caused can provide effective control of Swede midge on seed- damage to Þeld-planted spring canola plants in On- lings at the early stage of insect occurrence. tario, Canada (http://www.inspection.gc.ca/english/ During the early phase of regional colonization by plaveg/protect/rmd/rmd-08-03e.shtml). Damage to Swede midge in Canada, Þeld trials conducted in 2001 canola is dependent upon crop phenology at time of and 2002 indicated that foliar insecticide applications Swede midge infestation and may range from slight to were effective in keeping damage within marketable severe (R.H.H., unpublished data). Because of the limits for cabbage and broccoli plants (Hallett et al. proximity of infested areas in Canada, Ͼ8,498 ha 2009a). However by 2005 and 2006, insecticide treat- (Ͼ21,000 acres) of crucifers (cabbage, broccoli, cau- ments were rarely able to maintain damage levels lißower, Chinese cabbage, and Brussels sprouts) within marketable limits (i.e., damage rating Ͻ1). This grown in the northeastern United States, a value ex- was in spite of the fact that early season applications ceeding US$100 million, is particularly vulnerable to (e.g., seed treatments, greenhouse plug tray drenches, invasion by this pest. Swede midge has become es- band sprays, or a combination) of neonicotinoid in- tablished in parts of New York state, which may fur- secticides (e.g., imidacloprid, acetamiprid, and cloth- ther serve as a source for movement to other states in ianidin) could provide control of Swede midge for the United States (Kikkert et al. 2006). Substantial several weeks (Hallett et al. 2009a). Low efÞcacy of economic losses caused by Swede midge are likely to synthetic insecticides under Þeld conditions in some occur if effective control strategies are not developed areas of Canada was initially assumed to be associated and implemented. with insecticide resistance, spray coverage, and den- sity-dependent efÞcacy of insecticide applications (e.g., high Swede midge population pressure resulting Control Strategies in a higher number of survivors). Parallel comparison Since the Þrst identiÞcation of Swede midge in studies between a Þeld-collected Canadian Swede North America in 2000 (Hallett and Heal 2001), dif- midge population and a laboratory strain originating ferent control strategies, such as insecticidal control, from Switzerland indicated that both strains had sim- cultural control, host plant resistance, and biological ilar susceptibilities to seven common insecticides, control, have been extensively tested under labora- which suggested that insecticide resistance would not tory, greenhouse, and open Þeld conditions by re- be a key reason for the insecticide control failure in searchers in Canada and the United States. the Þeld (Hallett et al. 2009a). In addition, no signif- Insecticidal Control. One of the Þrst actions taken icant density-dependent and spray coverage effects for an immediate control of Swede midge and pre- were found in laboratory studies with Swede midge vention of further spread in North America was to occurring on caulißower plants when acetamiprid was screen synthetic insecticides labeled for crucifer veg- used as a foliar spray; however, the insecticide treat- June 2011 CHEN ET AL.: SWEDE MIDGE IN NORTH AMERICA 713 ment was only able to control Swede midge for9din of Swede midge, regardless of the presence of the Þelds (M.C. et al., unpublished data). Residual efÞ- cruciferous weeds. Similarly, a 3-yr Þeld survey dem- cacy trials indicate that foliar applications of regis- onstrated that cruciferous weeds may not be able to tered products have 4Ð10 d residual activity against maintain a Swede midge population in Þelds for Ͼ2yr Swede midge (R.H.H., unpublished data). Therefore, in the absence of preferred cultivated host plants, i.e., rather than high population pressure or insecticide cruciferous vegetables (Chen et al. 2009b). These lab- resistance, multiple and overlapping generations of oratory and Þeld studies suggest that crop rotation Swede midge, in conjunction with short residual ac- could be a powerful component of a Swede midge tivity of insecticides, might be the key reasons for management program. However, considering reports failure of control in Canada. Such results further in- that some proportion of the population may remain in dicate that proper insecticide application timing, and diapause for more than one winter (Rygg and Braekke multiple management tactics, will be crucial in achiev- 1980), and the availability of a wide range of crucif- ing acceptable control of Swede midge. A recent Þeld erous weeds, control of Swede midge by using crop test in Canada, undertaken where Swede midge pop- rotation needs to be evaluated in conjunction with ulations are quite high, showed that a drench of imi- weed management and Þeld sanitation efforts. dacloprid provided complete control for Ͼ25 d Other cultural practices include the use of barriers (A.M.S. et al., unpublished data). to prevent Swede midge adults from moving into Þelds Cultural Control. Along with insecticidal strategies, and managing planting and harvesting dates of cruci- control of Swede midge with cultural practices also fer crops. Previous studies have shown that 1.4-m-high has been investigated in North America. Because mesh fencing can reduce Swede midge damage to Swede midge spends its pupal stage in soil, the role of broccoli (Wyss and Daniel 2004). The use of a har- soil abiotic factors, such as soil type and moisture vestable barrier crop also may prove effective in re- content, have been explored for regulating population ducing host Þnding, and thus damage, by the Swede establishment and abundance. Under laboratory con- midge, and this is currently under investigation ditions, Chen and Shelton (2007) evaluated the im- (R.H.H. et al., unpublished data). Barriers may be pacts of six different types of soils collected in New effective in Þelds newly planted to crucifers and those York state on Swede midge pupation and emergence. in which crop rotation is practiced. However, it is not Swede midge emergence was signiÞcantly reduced in an appropriate approach in Þelds with continual cru- extremely wet or dry soils, regardless of soil type, cifer production, except potentially where the Þeld is suggesting that soil type alone is not a major factor planted and barrier erected after the emergence of the regulating Swede midge abundance. Optimal moisture overwintered generation of midges. The timing of content for Swede midge emergence varied from 25 to crop planting also may be manipulated to reduce the 75% among different soils. Chen and Shelton (2007) risk of damage by Swede midge. Planting only early found that most Swede midge pupated within the top season crucifer crops could reduce damage on plants 1cmofthesoilsandthattheadditionofϾ5cmofsoil at the early life stages when plants are most vulnerable cover could effectively reduce the emergence number to Swede midge. Broccoli production in some parts of and delay the time of emergence in laboratory studies. Sweden was reportedly abandoned as a result of high However, tillage was not found to reduce emergence Swede midge infestations in the late 1970s (B. Jonsson, in the Þeld, probably because tillage did not bury all personal communication). Although broccoli is sus- pupae at the maximum tillage depth but distributed ceptible to Swede midge infestation in all growing them throughout the soil column (R.H.H. et al., un- stages, crop phenology is an important factor in the published data). In addition, tillage probably in- severity of damage occurring. Susceptibility of cab- creased the depth to which precipitation could per- bage to Swede midge declines once heading has oc- colate and loosened the soil column, thus stimulating curred. Winter canola (i.e., fall-planted) and early emergence and facilitating the movement of pupae to planted spring canola experience little damage com- the surface (R.H.H. et al., unpublished data). pared with late planted spring canola (R.H.H. and Because crop rotation has proven to be effective to M. K. Sears, unpublished data). control other midge species in different crop systems Host Plant Resistance. Host plant resistance is a (Golightly and Woodville 1974, Faheemah and Sulai- major component of IPM and has been successfully man 1990), rotation also has suggested for Swede used for insect pest management (Stoner 1992, Smith midge (Taylor 1912, Rygg and Braekke 1980, Theunis- 2005). Swede midge feeds on a wide range of crucif- sen et al. 1997, ISMTF 2005). However, there were no erous plants (Barnes 1946; Stokes 1953a,b; Darvas et al. published data to validate this claim. Under conÞned 2000; Hallett 2007; Chen et al. 2009b), but these plants laboratory conditions, Chen et al. (2009a) evaluated display different degrees of susceptibility. The relative the control effects of 11 simulated caulißowerÐsweet resistance and susceptibility of broccoli, cabbage, cau- corn (Zea mays L.) and caulißowerÐkidney bean lißower, and Brussels sprouts to Swede midge was (Phaseolis vulgaris L.) crop rotation systems on Swede investigated in a 3-yr Ontario Þeld study conducted midge, also taking into account the potential role of shortly after the discovery of Swede midge (Hallett cruciferous weeds. The effectiveness of one cycle of 2007). In general, broccoli cultivars experienced more nonhost crop rotation could be reduced if cruciferous severe damage than cabbage, caulißower, and Brussels weeds were present; however, two consecutive cycles sprouts. ÔParagonÕ, ÔEurekaÕ, and ÔPackmanÕ broccoli of nonhost plant crop rotations provided full control (Stokes Seed Ltd., St. Catherines, ON, Canada) were 714 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3 highly susceptible to the Swede midge, whereas ÔTri- Challenges and Conclusions athlonÕ and ÔRegalÕ broccoli showed reduced suscep- Once well-established in an area, the Swede midge tibility to damage and slower development of damage is virtually impossible to eradicate. Consequently, the symptoms. No differences in damage severity and pro- Canadian Food Inspection Agency and USDA gression of damage symptoms were found between Plant Health Inspection Agency both changed the normal and red cultivars of cabbage and caulißower. status of Swede midge from a regulated or reportable/ In two different planting patterns (monoculture and actionable pest to a deregulated or nonreportable/ mixture planting), 23 common cultivars of cabbage, nonactionable pest in April 2009 (CFIA 2009, http:// broccoli, caulißower, and mustard (Harris Seeds, blogs.cce.cornell.edu/cvp/archives/63). Based on our Rochester, NY) in the northeastern United States research and that of others, it is evident that no single were evaluated for host plant tolerance to Swede control strategy is able to provide complete control of midge both under laboratory and Þeld conditions in Swede midge. Only an IPM approach that incorpo- New York State. Caulißower cultivar ÔAmerican-TakaÕ rates cultural techniques, sound chemical practices, was the most susceptible under both planting patterns, and use of potentially effective biological control while cabbage ÔLectroÕ and Mustard ÔGreenwaveÕ agents, could possibly suppress Swede midge popula- were the most tolerant (Chen et al. unpublished data). tions below economic thresholds. Although we have Such information could be used for identifying po- learned much about Swede midge in the last 10 yr in tential sources of resistance to the Swede midge, im- North America, much remains unclear and more chal- proving vegetable breeding programs, and providing lenges are yet to come. growers with planting recommendations. Based on our experience, the Swede midge has a Biological Control. Invasive species are generally scattered distribution in the United States and seems believed to be more successful in the new region of to have had many separate introduction points. Al- invasion than in their native region due in part to the though no longer considered a quarantine pest in absence of coevolved natural enemies, resulting in Canada or the United States, preventing introductions lower age speciÞc mortality of the invasive species to new regions within these countries is critical to (Drake 2003). In Europe, its region of origin, Swede minimizing the potential future impact of Swede midge larvae are parasitized by Pirene eximia Haliday midge. Further research is needed to evaluate the (Hymenoptera: Chalcididae) (Bovien and Knudsen impacts of various invasion paths (such as soil, trans- 1950), Synopeas sp. (Hymenoptera: Platygasteridae) plant movement, wind, adult ßight) on population (Rogerson 1963, Readshaw 1966, 1968), and Platy- establishment and vegetable production. Although a gaster sp. (Hymenoptera: Platygasteridae) (Readshaw considerable amount of research on understanding 1966). As a relatively new invader to North America, Swede midge biology and exploring management no specialized natural enemies have been found in the strategies has been done in Europe with European region (Corlay et al. 2007), although Medetera sp. strains (Hornig 1953; Readshaw 1966, 1968; Thygesen (Diptera: Dolichopodidae) reportedly prey on Swede 1966; Rygg and Braekke 1980; den Ouden et al. 1987; midge adults (Goodfellow 2005). Corlay et al. (2007) Theunissen et al. 1997; Wyss and Daniel 2004), genetic found that polyphagous coccinellid predators Harmo- variances between North America strains and Euro- nia axyridis (Pallas) and Coccinella septempunctata L. pean strains need investigation to track the regional could prey on Swede midge larvae under laboratory infestation patterns of speciÞc strains. Inconsistent conditions with the former showing a higher voracity control efÞcacies between laboratory and Þeld con- than the latter. However, it is unlikely that lady beetles ditions on Swede midge by using synthetic insecti- would forage for Swede midge larvae living between cides and soil manipulation require more research to tightly compressed tissue of whole plants. The ento- characterize the roles of various biotic and abiotic mopathogenic nematode Heterorhabditis bacterio- factors affecting Swede midge control, such as differ- phora, at an inoculation rate of 1,000 infective juve- ent emergence phenotypes (Hallett et al. 2009b), in- niles per larva, caused 90Ð100% mortality to Swede secticide deployment methods, and residual insecti- midge larvae in loam, sandy loam, clay, and muck soils cide activity. (Corlay et al. 2007). The soil bacterium Bacillus thu- Swede midge is becoming an increasingly important ringiensis subsp. israeliensis, when used as foliar spray pest of crucifers in North America, and its control has in the laboratory, caused signiÞcant mortality to proven difÞcult. It is imperative that better manage- Swede midge larvae on infested broccoli plants (Wu ment strategies be implemented while it is still a rel- et al. 2006), which is in agreement with other studies atively new pest, including novel insecticides and bio- in which the B. t. israeliensis strain showed activity to technology. A B. t. israeliensis strain showed activity dipteran species (Oestergaard et al. 2007). Vegetable against Swede midge (Wu et al. 2006) and other Bt cropping systems harbor great diversities of biological insecticidal proteins (e.g., Cry4Ba and Cry11Ba) are control agents (Word-Burkness et al. 2007) that may currently being tested at Cornell University. These be helpful for overall suppression of Swede midge. proteins could be used for generating Swede midge- However, to economically and sustainably control resistant genetically modiÞed crucifers, which may be Swede midge, further study is needed to identify and able to address the cryptic feeding behavior of Swede use more effective natural enemies of the pest. midge larvae. June 2011 CHEN ET AL.: SWEDE MIDGE IN NORTH AMERICA 715

Acknowledgments ual of Palaearctic Diptera. Science Herald, Budapest, Hungary. This work was partly supported by grants to A.M.S. from den Ouden, H., J. Theunissen, and A. M. Shelton. 1987. USDA Pest Management Alternatives Program and New Prevention of plant injury by cabbage gall midge (Con- York State Department of Agriculture and Markets, and New tarinia nasturtii Kieffer) and onion thrips (Thrips tabaci York State IPM program and to R.H.H. from the Ontario Lindemann) using emulsions of polyisobutylene. J. Appl. Ministry of Agriculture, Food & Rural Affairs (OMAFRA)Ð Entomol. 104: 313Ð331. University of Guelph Plants Program, OMAFRA New Direc- Drake, J. M. 2003. The paradox of the parasites: implications tions Research Fund, the Fresh Vegetable Growers of On- for biological invasion. Proc. R. Soc. London Ser. B 270: tario, and the Canola Council of Canada. S133ÐS135. Eastman, C., R. Barrido, H. Oloumi-Sadeghi, C. Hoy, J. Wy- man, J. Palumbo, G. Liebee and B. Flood. 2005. Cab- References Cited bage, broccoli and caulißower, pp. 157Ð174. In R. Foster and B. Flood [eds.], Vegetable insect management. Meis- Allen, J., H. Fraser, and R. H. Hallett. 2008. The Swede ter Media Worldwide, Willoughby, OH. midgeÑa pest of crucifer crops. OMAFRA Factsheet, Faheemah, S., and M. Sulaiman. 1990. The cucurbit gall Agdex#: 625/252. (http://www.omafra.gov.on.ca/english/ midge, Lasioptera chichindae Grover (Diptera: Ceci- crops/facts/08-007.htm). domyiidae) and its possible control. Trop. Agric. 146: Andaloro, J. T., C. W. Hoy, K. B. Rose, and A. M. Shelton. 67Ð72. 1983. Evaluation of insecticide usage in the New York Finch, S., and A. R. Thompson. 1992. Pests of cruciferous processing cabbage pest management program. J. Econ. crops, pp. 87Ð138. In R. G. McKinlay [ed.], Vegetable crop Entomol. 76: 1121Ð1124. pests. CRC, Boca Raton, FL. Bardner, H. M., C. A. Edwards, M. K. Arnold, and J. P. Frey, J. E., B. Frey, and R. Baur. 2004. Molecular identiÞ- Rogerson. 1971. The symptoms of attack by swede cation of the swede midge (Diptera: Cecidomyiidae). midge (Contarinia nasturtii)andtheeffectsonyieldof Can. Entomol. 136: 771Ð780. Swedes. Entomol. Exp. Appl. 14: 223Ð233. Golightly, W. H., and H. C. Woodville. 1974. Studies of Barnes, H. F. 1946. Gall midges of economic importance, recent outbreaks of saddle gall midge. Ann. Appl. Biol. 77: vol. I: Gall midges of root and vegetable crops. Crosby, 97Ð101. Lockwood & Son, London, United Kingdom. Boddum, T., N. Skals, M. Wiren, R. Baur, S. Rauscher, and Y. Goodfellow, S. A. 2005. Population dynamics and predictive Hillbur. 2009. Optimisation of the pheromone blend of modeling of the swede midge, Contarinia nasturtii (Kief- the swede midge, Contarinia naturtii, for monitoring. Pest fer). M.S. Thesis, University of Guelph, Guelph, ON, Manag. Sci. 65: 851Ð856. Canada. Bovien, P., and P. Knudsen. 1950. Krusesygegalmyggen Hallett, R. H. 2007. Host plant susceptibility to the Swede (Contarinia nasturtii Kieff.), dens biologi og bekaem- midge, Contarinia nasturtii (Diptera: Cecidomyiidae). J. pelse. Tid. Planteaul. 53: 235Ð257. Econ. Entomol. 100: 1335Ð1343. [CFIA] Canadian Food Inspection Agency. 2008. 2008 Hallett, R. H., and J. D. Heal. 2001. First Nearctic record of Plant protection survey report. CFIA Plant Health the swede midge, Contarinia nasturtii (Kieffer) (Diptera: Division. (http://www.inspection.gc.ca/english/plaveg/ Cecidomyiidae), a pest of cruciferous crops in Europe. pestrava/surv/sit2008e.shtml#connas). Can. Entomol. 133: 713Ð715. [CFIA] Canadian Food Inspection Agency. 2009. RMD-08- Hallett, R. H., S. A. Goodfellow, and J. D. Heal. 2007. Mon- 03: review of the pest status of the swede midge (Con- itoring and detection of the swede midge (Diptera: Ce- tarinia nasturtii)inCanada.CFIAPlantHealthDivision. cidomyiidae). Can. Entomol. 139: 700Ð712. (http://www.inspection.gc.ca/english/plaveg/protect/ Hallett, R. H., M. Chen, M. K. Sears, and A. M. Shelton. rmd/rmd-08-03e.shtml). 2009a. Insecticide management strategies for control of Chen, M., and A. M. Shelton. 2007. Impact of soil type, Swede midge (Diptera: Cecidomyiidae) on cole crops. J. moisture and depth on Swede midge pupation and emer- Econ. Entomol. 102: 2241Ð2254. gence. Environ. Entomol. 36: 1349Ð1355. Hallett, R. H., S. A. Goodfellow, R. M. Weiss, and O. Olfert. Chen, M., J. Z. Zhao, and A. M. Shelton. 2007. Control of 2009b. MidgEmerge, a new predictive tool, indicates the Contarinia nasturtii (Diptera: Cecidomyiidae) by foliar presence of multiple emergence phenotypes of the over- sprays of acetamiprid on caulißower transplants. Crop wintered generation of swede midge. Entomol. Exp. Appl. Prot. 26: 1574Ð1578. 130: 81Ð97. Chen, M., W. W. Li, and A. M. Shelton. 2009a. Simulated Harris, K. M. 1966. Gall midges genera of economic impor- crop rotation systems control swede midge, Contarinia tance (Diptera: Cecidomyiidae). Part I: Introduction and nasturtii. Entomol. Exp. Appl. 133: 84Ð91. subfamily Cecidomyiinae; supertribe Cecidomyiidi. Chen, M., A. M. Shelton, P. Wang, C. A. Hoepting, W. C. Trans. R. Entomol. Soc. Lond. 118: 313Ð358. Kain, and D. C. Brainard. 2009b. Occurrence of the new Hatchett, J. H., G. L. Kreitner, and R. J. Elzinga. 1990. Lar- invasive insect, Contarinia nasturtii, on cruciferous val mouthparts and feeding mechanism of the Hessian ßy weeds. J. Econ. Entomol. 102: 115Ð120. (Diptera: Cecidomyiidae). Ann. Entomol. Soc. Am. 83: Corlay, F., and G. Boivin. 2008. Seasonal development of an 1137Ð1147. invasive exotic species, Contarinia nasturtii (Diptera: Ce- Hillbur, Y., M. Celander, R. Baur, S. Rauscher, J. Haftmann, cidomyiidae), in Quebec. Environ. Entomol. 37: 907Ð913. S. Franke, and W. Francke. 2005. IdentiÞcation of the Corlay, F., G. Boivin, and G. Belair. 2007. EfÞciency of sex pheromone of the swede midge, Contarinia nasturtii. natural enemies against the swede midge Contarinia nas- J. Chem. Ecol. 31: 1807Ð1828. turtii (Diptera: Cecidomyiidae), a new invasive species in Hornig, H. 1953. The signiÞcance of C. nasturtii for swede North America. Biol. Control 43: 195Ð201. cultivation with special reference to conditions in Darvas, B., M. Skuhrava, and A. Andersen. 2000. Agricul- Schleswig Holstein. Z. Ang. Entomol. 35: 271Ð318. tural and dipteran pests of the Palaearctic region, vol. 1. [ISMTF] International Swede Midge Task Force. 2005. In L. Papp and B. Darvas [eds.], Contributions to a man- 2005 Interim best management practices to control the 716 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3

swede midge (Contarinia nasturtii Kieffer). (http://www. Shelton, A. M., M. K. Sears, J. A. Wyman, and T. C. Quick. omafra.gov.on.ca/english/crops/facts/bmp_swedemidge. 1983. Comparison of action thresholds for lepidopterous htm). larvae on fresh market cabbage. J. Econ. Entomol. 76: Jones, S., and A. E. Luchsinger. 1979. Introduction to plant 196Ð199. systematics. McGraw-Hill, New York. Shelton, A. M., W. T. Wilsey, and M. A. Schmaedick. 1998. Jyoti, J., A. M. Shelton, and J. Barnard. 2003. Evaluation of Management of onion thrips on cabbage using plant re- degree-day and Julian-day logistic models predicting cab- sistance and insecticides. J. Econ. Entomol. 91: 329Ð333. bage maggot (Diptera: Anthomyiidae) emergence and Smith, C. M. 2005. Plant resistance to : molecular ßight in upstate New York. J. Entomol. Sci. 38: 525Ð532. and conventional approaches. Springer, Dordrecht, The Kikkert, J. R., C. A. Hoepting, Q. J. Wu, P. Wang, R. Baur, and Netherlands. A. M. Shelton. 2006. Detection of Contarinia nasturtii Statistics Canada. 2010. Field and special crops seeded area. (Diptera; Cecidomyiidae) in New York, a new pest of (http://www.statcan.gc.ca/start-debut-eng.html). cruciferous plants in the United States. J. Econ. Entomol. Stokes, B. M. 1953a. Biological investigations into the valid- 99: 1310Ð1315. ity of Contarinia species living on the Cruciferae, with Mika, A. M., R. M. Weiss, O. Oflert, R. H. Hallett, and J. A. special reference to the swede midge, Contarinia nasturtii Newman. 2008. Will climate change be beneÞcial or det- (Kieffer). Ann. Appl. Biol. 40: 726Ð741. rimental to the invasive swede midge in North America? Stokes, B. M. 1953b. The host plant range of the swede Contrasting predictions using climate projections from midge (Contarinia nasturtii Kieffer) with special refer- different general circulation models. Global Change Biol. ence to types of plant damage. Tijdschr. Plantenziekten. 14: 1721Ð1733. 59: 82Ð90. Oestergaard, J., R. Ehlers, A. C. Martinez-Ramirez, and M. D. Stoner, K. A. 1992. Bibliography of plant resistance to ar- Real. 2007. Binding of Cyt1Aa and Cry11Aa toxins of thropods in vegetables, 1977Ð1991. Phytoparasitica 20: Bacillus thuringiensis serovar israelensis to brush border 125Ð180. membrane vesicles of Tipula paludosa (Diptera: Nema- Taylor, T. H. 1912. Cabbage-top in swedes. The University of Leeds and the Yorkshire Council for Agricultural Ed- tocera) and Subsequent Pore Formation. Appl. Environ. ucation 82: 1Ð21. Microbiol. 73: 3623Ð3629. Theunissen, J., H. den Ouden, and G. Schelling. 1997. Can Olfert, O., R. Hallett, R. M. Weiss, J. Soroka, and S. Good- the cabbage gall midge, Contarinia nasturtii (Diptera, fellow. 2006. Potential distribution and relative abun- Cecidomyiidae) be controlled by host plant deprivation? dance of swede midge, Contarinia nasturtii, an invasive Meded. Fac. Landbouwk. Toegep. Biol. Wet. Univ. Gent pest in Canada. Entomol. Exp. Appl. 120: 221Ð228. 62: 617Ð622. [OMAFRA] Ontario Ministry of Agriculture, Food and Ru- Thygesen, T. 1966. The swede midge (C. nasturtii). Inves- ral Affairs. 1996. Integrated pest management for cru- tigations on bionomics and economic importance, to- cifers. Publication 701. Ontario Ministry of Agriculture, gether with control experiments. Tidsskr. Planteavl. 70: Food & Rural Affairs, Guelph, ON, Canada. 179Ð197. Readshaw, J. L. 1961. The biology and ecology of the swede [USDA–ERS] U.S. Department of Agriculture–Economic midge, Contarinia nasturtii (Kieffer) (Diptera: Ceci- Research Service. 2010. Homepage. (http://www.ers. domyiidae). Ph.D. dissertation, University of Durham, usda.gov). Durham, United Kingdom. Word-Burkness, S. J., W. D. Hutchison, J. C. Lee, R. L. Hines, Readshaw, J. L. 1966. The ecology of the swede midge, Con- P. C. Bolin, and G. E. Heimpel. 2007. Along-termsurvey tarinia nasturtii (Kieffer) (Diptera: Cecidomyiidae). I. of parasitoid species composition and parasitism of Life-history and inßuence of temperature and moisture Trichoplusia ni (Lepidoptera: Noctuidae), Plutella xylos- on development. Bull. Entomol. Res. 56: 685Ð700. tella (Lepidoptera: Plutellidae), and Pieris rapae (Lepi- Readshaw, J. L. 1968. Damage to swedes by the swede doptera: Pieridae) in Minnesota cabbage. J. Entomol. Sci. midge, Contarinia nasturtii (Kieff.), and a possible 40: 211Ð221. method of cultural control. Bull. Entomol. Res. 58: 25Ð29. Wu, Q. J., J. Z. Zhao, A. G. Taylor, and A. M. Shelton. 2006. Reiners, S., and C. H. Petzoldt. 2010. Integrated crop and pest Evaluation of insecticides and application methods management guidelines for commercial vegetable produc- against Contarinia nasturtii (Diptera: Cecidomyiidae), a tion. Cornell Cooperative Extension, Ithaca, NY. (http:// new invasive insect pest in the United States. J. Econ. www.nysaes.cornell.edu/recommends/1frameset.html). Entomol. 99: 117Ð122. Rogerson, J. P. 1963. Swede midge on two Northumberland Wyss, E., and C. Daniel. 2004. The effect of exclusion fences farms, 1959Ð1961. Plant Pathol. 12: 161Ð171. on the colonization of broccoli and by the swede Rygg, T. D., and H. P. Braekke. 1980. Swede midge (Con- midge, Contarinia nasturtii (Diptera: Cecidomyiidae). tarinia nasturtii Kieffer) (Diptera, Cecidomyiidae): In- Mitt. Dtsch. Ges. Allg. Ang. Entomol. 14: 387Ð390. vestigations on biology, symptoms of attack and effects on yield. Meld. Norg. Landbruk. 59: 1Ð9. Received 24 October 2010; accepted 28 December 2010.